The Design And Implementation On Parallel Query Acceleratror

With the continuous advancement of IT technology, the computer system in the national economy and people's daily life plays an increasingly important role, but also produced more and more date. According to wintercorp corporate investigations, the world's largest database every two years, triple the size, we have truly entered into the "sea of data" management era. Data on the size of the GB level from the past into the TB level are now close to PB level. Network monitoring, e-commerce, real-time data warehouses have become an important source of mass data. On the other hand, the rapid development of computing power and ever-increasing storage density makes it possible to mass data processing.OLTP is a traditional relational database in the main applications with large data transfers, but as the data transmission technology, the technology has seriously constrained the performance of information systems, although the stand-alone hardware configuration can be used to improve and optimize the performance parameters of the method of database systems on the system performance improvements, but the results are unsatisfactory. Therefore, the database clustering technology is becoming the main contents of the database field of study, based on multi-machine clusters for parallel processing of database-based high performance, strong usability and scalability in high-performance computing, mass data storage and processing, Web services, e-commerce areas such as playing a great role.The application of the data is generated at any time, because the system is huge, but the data production rate is high. These data need to save them in a timely manner and conduct statistical analysis to arrive at an effective information, requiring large amounts of data scanning and computing, the traditional single parallel database difficult to handle such a large scale of data.This parallel database cluster middleware ERAC is a non-shared (shared-nothing) architecture of parallel database software. It uses Oracle database as a data processing unit to write data in parallel a number of independent homogeneous database queries in parallel to read and process the data, which greatly improves the ability to store data in the system and shorten a single query completion time. By improving the hardware configuration of a single database node, and increasing the number of database nodes can achieve the data capacity and processing capacity expansion. ERAC is configured with separate loading and query services such as node, by configuring a reasonable number of service nodes can give full play to the database node in the storage and processing capacity, and improve the system's overall service capabilities. Load balancing will be distributed to the user requests a different middleware server. ERAC user can provide a similar JDBC programming interface and practical tools transparent, efficient access to parallel database system.